High-energy and low-cost membrane-free chlorine flow battery

The study addresses the need for low-cost, efficient, and scalable grid-scale energy storage to integrate intermittent renewables. While redox flow batteries are promising, current chemistries (notably all-vanadium and engineered organics) suffer from high material costs, limited energy density, and reliance on expensive ion-permeable membranes. The authors propose leveraging the abundant and inexpensive NaCl via the reversible Cl₂/Cl⁻ couple, which offers high theoretical capacity, fast kinetics, and low cost, to create a membrane-free chlorine flow battery by separating phases (organic Cl₂ solvent vs aqueous NaCl). The research goal is to demonstrate a high-efficiency, high-energy-density, low-cost, and membrane-free chlorine redox flow battery and elucidate the mechanisms enabling its performance.

The paper reviews limitations of incumbent RFB chemistries: all-vanadium systems face high chemical costs and low energy density; aqueous organic redox couples require tailored synthesis and still depend on costly ion-selective membranes. Polymer redox couples and semi-solid Li-ion flow systems mitigate crossover or increase energy density but introduce high viscosity, cost, and power limitations. Historically, Cl₂/Cl⁻ redox was used (e.g., in the 1884 La France airship), but prior chlorine-based systems suffered low Coulombic efficiency (due to Cl₂ dissolution in electrolyte) and large voltage hysteresis (wettability issues), and attempts such as Cl₂ graphite intercalation were unstable at room temperature. These gaps motivate re-examining chlorine chemistry with phase-separated storage to avoid crossover and membrane use.

Electrochemical cell configurations: A three-electrode concentric cell evaluates the Cl₂/Cl⁻ redox in saturated NaCl aqueous electrolyte (NaCl/H₂O), with a RuO₂–TiO₂ coated porous carbon working electrode (RuO₂–TiO₂@C), an activated carbon counter electrode, and an Ag/AgCl reference. CCl₄ is pumped through the working electrode; NaCl/H₂O flows through the interstitial space between electrodes, ensuring Cl₂ supply. Contact angle measurements (<90° for both CCl₄ and NaCl/H₂O on graphite) confirm wetting; volumetric occupancy in the porous electrode is about 66.2% CCl₄ and 33.8% NaCl/H₂O. The immiscibility of CCl₄ and NaCl/H₂O obviates an ion-permeable membrane. Electrochemical reactions: Positive electrode: 2Cl⁻ − 2e⁻ ↔ Cl₂ (E⁰ = 1.36 V vs NHE). For full-cell operation, the negative electrode is NaTi₂(PO₄)₃, operating at ~−0.5 V vs NHE with rapid and reversible Na⁺ insertion/extraction (validated by CV). Full chlorine flow battery (CFB): The counter is replaced by a NaTi₂(PO₄)₃ negative electrode to assemble a full cell. Flow rates typically Qaq = 0.02 mL/s (NaCl/H₂O) and Qorg = 0.002 mL/s (CCl₄). Galvanostatic charge/discharge is conducted across current densities, with state-of-charge normalized to 600 mAh reversible capacity at 10 mA/cm². Voltage efficiency is defined as discharge-to-charge voltage ratio; energy efficiency is voltage efficiency × Coulombic efficiency. Modeling: A steady-state model couples Nernst–Planck transport in the porous RuO₂–TiO₂@C electrode (cell width 0–1.0 mm) and NaCl/H₂O (1.0–4.0 mm) with Fickian diffusion in the Cl₂–CCl₄ phase (−2.0–0 mm). The negative electrode is imposed as a boundary at the far side of the aqueous phase. Simulations predict potential distributions and species concentrations (Cl⁻ and Cl₂) and are validated against measured cell voltages over varying flow rates and current densities. Materials synthesis: RuO₂/TiO₂ catalysts on activated carbon are prepared by impregnating RuCl₃ and titanium isopropoxide in isopropanol onto activated carbon, drying (90 °C), and annealing at 500 °C for 1 h in air. Carbon-coated NaTi₂(PO₄)₃ is synthesized from Na₂CO₃, NH₄H₂PO₄, and TiO₂ in 2 wt% PVA solution, dried and calcined at 900 °C (10 h, N₂), followed by thermal vapor deposition of carbon using toluene at 700 °C (2 h) with N₂ flow, and post-heat-treatment at 900 °C (2 h). Electrode fabrication and testing: Electrodes consist of active material (porous carbon or carbon-coated NaTi₂(PO₄)₃), carbon black, and PTFE binder (7:2:1) pressed onto titanium grid (10 MPa). CVs are collected with a CHI660B workstation; galvanostatic cycling is performed on an Arbin system. Characterization includes SEM, PXRD, Raman, and viscosity measurement (CANNON-FENSKE viscometer). Energy density calculation: Based on a 600 mAh cell at 1.8 V average (10 mA/cm²). Active volumes: 6.0 mL CCl₄, 2.0 mL NaCl/H₂O, and 0.592 mL NaTi₂(PO₄)₃ (5.0 g, density 2.96 g/mL). Cell-level energy density based on active materials is 125.7 Wh/L.

• Membrane-free configuration enabled by immiscible Cl₂ solvent (CCl₄) and aqueous NaCl eliminates crossover and membrane losses, maintaining phase separation within a wetted porous electrode.
• Coulombic efficiency improves from 8% (without CCl₄) to 97% (with CCl₄) due to preferential storage of Cl₂ in CCl₄ (three orders higher solubility vs NaCl/H₂O: 0.184 vs 0.0005 mole/mole).
• Cl₂–CCl₄ delivers 97 Ah/L volumetric capacity (600 mAh reversible capacity with 6.0 mL CCl₄), exceeding typical vanadium catholytes (22.6–43.1 Ah/L).
• Viscosity of the positive electrolyte is low and nearly constant; Cl₂–CCl₄ decreases from 0.894 to 0.819 mPa·s up to saturation, facilitating low pumping loss.
• Full-cell voltage efficiency: 93.6% at 10 mA/cm² and ~77% at 100 mA/cm²; peak power density reaches 325 mW/cm² at 350 mA/cm².
• Round-trip energy efficiency: 91% at 10 mA/cm²; cell-level energy density: 125.7 Wh/L (based on active materials).
• Modeling shows charge limited by Cl⁻ transport in aqueous phase and discharge limited by Cl₂ in organic phase; Cl⁻ gradients dominate due to lower diffusivity and smaller aqueous volume fraction, explaining asymmetric overpotentials.
• Potential loss across the aqueous phase is ~20 mV at 10 mA/cm² and ~250 mV at 100 mA/cm², over five times smaller than ion transport losses across Nafion membranes in comparable aqueous RFBs.
• Cycling stability demonstrated: stable voltage profiles at 20 mA/cm² and capacity retention over 500 cycles (with 600 mAh charge capacity).
• Alternative organic phases: mineral spirit yields 91.6 Ah/L at 20 °C with good wettability and low viscosity; candidates include heptane, octane, and tetradecane.
• Multivalent charge carriers possible: adding ZnCl₂ allows Zn metal negative electrode, increasing cell voltage to 1.9 V.
• Estimated active material cost for storage is ~ $5/kWh; membrane-free design removes a component typically >30% of system cost, indicating strong cost-competitiveness vs existing RFBs.

The work demonstrates that storing electrochemically generated chlorine in an immiscible, highly soluble organic phase (CCl₄ or mineral spirits) while conducting ionic transport in an aqueous NaCl phase enables a membrane-free redox flow architecture. This configuration suppresses chlorine crossover, enhances Coulombic efficiency, and reduces transport overpotentials otherwise imposed by ion-selective membranes. Fast Cl₂/Cl⁻ kinetics and excellent wetting of porous carbon maximize reactive surface area and mass transport, yielding high voltage efficiency and power density. Modeling clarifies that different limiting species (Cl⁻ during charge, Cl₂ during discharge) and their diffusivities/volume fractions govern overpotentials, and flow rate optimization can mitigate these gradients. The approach attains high round-trip efficiency and energy density with very low active material cost, directly addressing the central challenge of cost-effective, reliable, and scalable stationary storage. Safety considerations are addressed by leveraging closed-system designs and industrial chlorine handling practices.

A membrane-free chlorine flow battery using phase-separated Cl₂ storage in CCl₄ (or mineral spirit) and aqueous NaCl as the electrolyte achieves high energy efficiency (91%), high energy density (125.7 Wh/L), substantial power capability, and very low materials cost (~$5/kWh). The system benefits from fast Cl₂/Cl⁻ kinetics, superior wetting, low viscosity, and minimized transport loss without membranes. Modeling and experiments reveal mass-transport-controlled overpotentials and provide strategies (flow management) to further improve performance. The concept broadens RFB design space by enabling both anionic and cationic charge carriers and compatibility with multivalent ions (e.g., Zn), and suggests practical solvents beyond CCl₄. Future work may focus on solvent optimization (safer, cheaper, high-Cl₂-solubility), scaling of cell architecture, replacement or reduction of precious catalysts, system-level safety engineering, and long-term durability under practical cycling conditions.

• Use of CCl₄ serves as a proof-of-concept solvent; it has toxicity concerns, necessitating alternative solvents (e.g., mineral spirits) for practical deployment.
• Chlorine handling requires strict safety measures; while the system is closed and industrial practices exist, scaling introduces potential exposure and leak mitigation challenges.
• Performance asymmetry and overpotentials arise from mass transport limitations (Cl⁻ in aqueous phase during charge, Cl₂ in organic phase during discharge), which may constrain high-current operation without optimized flow and architecture.
• The demonstration is at small lab scale; system-level engineering (stack design, long-term sealing/permeation control, balance-of-plant) and extended lifetime testing beyond 500 cycles are needed.
• Catalysts (RuO₂–TiO₂) contribute cost; although partial substitution by cheaper oxides is proposed, catalyst optimization remains to further reduce cost and ensure durability.